matrix multi

[Aries2003]

i hav a doubt as u can see when i used the below commands when i try to save the matrix multi in same tensor (that is tensor1 in this case) it breaks the first rule of MM that is that the inner dimension should of same size
This gives an output which is wrong right?
but if i tranpose it its proper ...My question is it shouldnt work in the first place by rule one right i talking about the first code which gives different ans

tensor1= torch.mm(tensor1,tensor2)
tensor1

this gives ouput:
tensor([[0.0935, 2.4341, 1.7364, 2.7921, 2.6920, 2.7091, 1.1497],
        [0.0475, 1.2374, 0.8827, 1.4194, 1.3685, 1.3772, 0.5844],
        [0.0577, 1.5026, 1.0718, 1.7235, 1.6617, 1.6723, 0.7097],
        [0.0434, 1.1304, 0.8063, 1.2966, 1.2501, 1.2581, 0.5339],
        [0.0894, 2.3274, 1.6602, 2.6696, 2.5739, 2.5903, 1.0992],
        [0.0583, 1.5180, 1.0829, 1.7413, 1.6788, 1.6895, 0.7170],
        [0.0769, 2.0029, 1.4287, 2.2974, 2.2150, 2.2291, 0.9460]])

tensor= torch.mm(tensor1,tensor2.T)
tensor
where as this gives :
tensor([[11.1213],
        [ 5.6535],
        [ 6.8650],
        [ 5.1646],
        [10.6335],
        [ 6.9357],
        [ 9.1508]])


tensor1= torch.mm(tensor1,tensor2.T)
tensor1

tensor([[11.1213],
        [ 5.6535],
        [ 6.8650],
        [ 5.1646],
        [10.6335],
        [ 6.9357],
        [ 9.1508]])

[LuluW8071]

You need to clarify more by giving the original tensor1 and tensor2 matrix values